Alaskan Yelloweye Rockfish fecundity revealed through an automated egg count and digital imagery method

Spawning stock biomass (SSB) is often used as an index for reproductive potential (RP) in fisheries stock assessments. This method assumes that mature female biomass is proportional to total egg production and implies that (1) the fecundity–length relationship follows a cubic function or (2) relative fecundity is constant. For many marine fishes, adequate fecundity estimates to evaluate these relationships are lacking. This study estimated fecundity and fecundity relationships for Yelloweye Rockfish Sebastes ruberrimus and evaluated an automated method of counting eggs and larvae. We collected Yelloweye Rockfish ovaries (N = 90) from the northern Gulf of Alaska, including Prince William Sound, Alaska, during 2018–2019 and used the gravimetric method and image analysis software to count eggs from digital camera images. To evaluate the speed, accuracy, and precision of the automated counting procedure, one-third of the gravimetric samples were also manually counted. Image analysis software was approximately four times faster but equally accurate and precise for fecundity estimates relative to manual counts. Fecundity ranged from 53,249 to 3.052 × 10⁶ eggs (mean ± SD = 896,762 ± 699,504 eggs), and relative fecundity increased with female FL and ranged from 68 to 435 eggs/g of body weight (mean ± SD = 226 ± 87 eggs/g). The use of SSB for Yelloweye Rockfish stock assessment could underestimate the contribution to egg production by larger (>5.6-kg) females, overestimate the contribution by smaller females, and lead to biased biological reference points. This study provides critical information to more realistically model RP and improve stock assessment inputs for the development of harvest control rules for Yelloweye Rockfish. Additionally, the use of image analysis software to count eggs in digital images proved to be an effective fecundity estimation method that could be applied to other highly fecund fish species for which the time demand of manual counting methods would be prohibitive.